The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

Geothermal Reservoir Simulation Dataset

A dataset of 1,000 synthetic geothermal reservoir simulations generated using implicit finite-difference solvers for coupled heat conduction and Darcy flow. Parameter ranges are calibrated to Indonesian vapor-dominated geothermal systems (Kamojang, Wayang Windu, Darajat).

This dataset was used to train GeoForce-CNN v1.1, a physics-informed surrogate model that achieves R²=0.994 for temperature and R²=0.997 for pressure prediction.

Dataset Description

Each of the 1,000 scenarios represents a 2D geothermal reservoir simulation over 20 years, outputting temperature and pressure fields at 5 timesteps on a 32×32 grid.

Per-Scenario Contents (`.npz` format)

Array	Shape	Unit	Description
`temperature`	(5, 32, 32)	°C	Temperature field at years 4, 8, 12, 16, 20
`pressure`	(5, 32, 32)	Pa	Pressure field at years 4, 8, 12, 16, 20
`well_locations`	(n_wells, 2)	grid indices	Production well coordinates
`param_keys`	(5,) or (6,)	—	Parameter names
`param_values`	(5,) or (6,)	mixed	Parameter values

Parameter Ranges (Latin Hypercube Sampling)

Parameter	Min	Max	Unit	Reference
Base temperature	180	320	°C	Kamojang (245°C), Wayang Windu (270°C), Darajat (250°C)
Base pressure	5	25	MPa	Hydrostatic at 500–2500 m depth
Log₁₀ permeability	-16	-12	log₁₀(m²)	Fractured volcanic rock
Porosity	0.01	0.15	fraction	Dense to moderately porous andesite
Depth	800	2500	m	Shallow to deep Indonesian reservoirs
Well count	1	5	count	Small to medium field

Grid Specification

Property	Value
Grid dimensions	32 × 32 cells
Domain size	1,000 × 1,000 m
Cell size	31.25 m
Simulation duration	20 years
Output timesteps	5 (years 4, 8, 12, 16, 20)

Physics Model

The simulator solves coupled PDEs using backward Euler (implicit) time-stepping with second-order central differences:

Heat equation:

ρ_eff · Cp_eff · ∂T/∂t = ∇·(k_thermal · ∇T) + Q_wells

Darcy flow (mass balance):

φ · β · ∂P/∂t = ∇·((k_perm / μ) · ∇P) + q_wells

Fixed physical constants:

Parameter	Value	Unit
Rock density	2700	kg/m³
Rock specific heat	900	J/(kg·K)
Water density	1000	kg/m³
Water specific heat	4186	J/(kg·K)
Thermal conductivity	2.5	W/(m·K)
Water viscosity	3×10⁻⁴	Pa·s
Fluid compressibility	4.5×10⁻¹⁰	1/Pa

Data Splits

Deterministic split using seed=42:

Split	Scenarios	Indices File
Train	800	`train_indices.npy`
Validation	100	`val_indices.npy`
Test	100	`test_indices.npy`

Quick Start

import numpy as np

# Load a single scenario
data = np.load("scenarios/scenario_00042.npz")
temperature = data["temperature"]  # (5, 32, 32) in °C
pressure = data["pressure"]        # (5, 32, 32) in Pa
wells = data["well_locations"]     # (n_wells, 2) grid coords
params = dict(zip(data["param_keys"], data["param_values"]))

print(f"Base temp: {params.get('base_temperature', 'N/A')}°C")
print(f"Year 20 avg temp: {temperature[4].mean():.1f}°C")
print(f"Wells: {len(wells)} at {wells.tolist()}")

Files

geothermal-reservoir-dataset/
├── README.md
├── scenarios/
│   ├── scenario_00000.npz
│   ├── scenario_00001.npz
│   ├── ...
│   └── scenario_00999.npz
├── train_indices.npy    (800 indices)
├── val_indices.npy      (100 indices)
└── test_indices.npy     (100 indices)

Citation

@dataset{forcex_geothermal_2026,
  title={Geothermal Reservoir Simulation Dataset},
  author={Riupassa, Robi Dany},
  year={2026},
  publisher={ForceX AI},
  url={https://huggingface.co/datasets/ForceX-AI/geothermal-reservoir-dataset},
  note={1,000 synthetic simulations calibrated to Indonesian geothermal fields}
}

References

Pruess, K., "TOUGH2: A General-Purpose Numerical Simulator for Multiphase Fluid and Heat Flow," LBL-29400, 1991.
Darma, S. et al., "Geothermal Energy Update: Development and Utilization in Indonesia," WGC 2010.
Hochstein, M.P., Sudarman, S., "History of Geothermal Exploration in Indonesia from 1970 to 2000," Geothermics, 2008.
Saptadji, N.M., "Reservoir Engineering of Geothermal Systems in Indonesia," IOP Conf. Series, 2017.